Face Generation¶
In this project, you'll use generative adversarial networks to generate new images of faces.
Get the Data¶
You'll be using two datasets in this project:
- MNIST
- CelebA
Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.
If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".
data_dir = './data'
# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper
helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)
show_n_images = 25
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')
CelebA¶
The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.
show_n_images = 25
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))
Preprocess the Data¶
Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.
The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).
Build the Neural Network¶
You'll build the components necessary to build a GANs by implementing the following functions below:
model_inputsdiscriminatorgeneratormodel_lossmodel_opttrain
Check the Version of TensorFlow and Access to GPU¶
This will check to make sure you have the correct version of TensorFlow and access to a GPU
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf
# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer. You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))
# Check for a GPU
if not tf.test.gpu_device_name():
warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))
Input¶
Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:
- Real input images placeholder with rank 4 using
image_width,image_height, andimage_channels. - Z input placeholder with rank 2 using
z_dim. - Learning rate placeholder with rank 0.
Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)
import problem_unittests as tests
def model_inputs(image_width, image_height, image_channels, z_dim):
"""
Create the model inputs
:param image_width: The input image width
:param image_height: The input image height
:param image_channels: The number of image channels
:param z_dim: The dimension of Z
:return: Tuple of (tensor of real input images, tensor of z data, learning rate)
"""
# TODO: Implement Function
inputs_real = tf.placeholder(tf.float32, [None, image_width, image_height, image_channels], name='input_real')
inputs_z = tf.placeholder(tf.float32, [None, z_dim], name='input_z')
learning_rate = tf.placeholder(tf.float32)
return inputs_real, inputs_z, learning_rate
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)
Discriminator¶
Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).
def discriminator(images, reuse=False):
"""
Create the discriminator network
:param images: Tensor of input image(s)
:param reuse: Boolean if the weights should be reused
:return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
"""
# TODO: Implement Function
# I'll try the suggested value(suggested in the DCGAN paper)
# of stddev 0.02 for initializing weights
# "Unsupervised Representation Learning with
# Deep Convolutional Generative Adversarial Networks":
# https://arxiv.org/pdf/1511.06434.pdf
# p.3, "4. Details of Adversarial Training"
kinit = tf.random_normal_initializer(stddev=0.02)
with tf.variable_scope("discriminator", reuse=reuse):
# layer1 with 28x28 size input
# this layer should not have any batch normalization process
# Also, after many tries, I decided to put a large number of filters
# for both discriminator and generator
x1 = tf.layers.conv2d(images, 128, 5, strides=2, padding='SAME',
kernel_initializer=kinit)
# Leaky ReLU for discriminator layers, with alpha value of 0.2
# as suggested in the DCGAN paper
relu1 = tf.maximum(0.2 * x1, x1)
# padding is 'same'
# so ceil(28 / 2) = 14
# now the shape is 14x14x128
# layer2
x2 = tf.layers.conv2d(relu1, 256, 5, strides=2, padding='SAME',
kernel_initializer=kinit)
# now batch normalization
bn2 = tf.layers.batch_normalization(x2, training=True)
relu2 = tf.maximum(0.2 * bn2, bn2)
# now the shape is 7x7x256
# layer3
x3 = tf.layers.conv2d(relu2, 512, 5, strides=2, padding='SAME',
kernel_initializer=kinit)
bn3 = tf.layers.batch_normalization(x3, training=True)
relu3 = tf.maximum(0.2 * bn3, bn3)
# now the shape is 4x4x512
# Also after many tries, I noticed that to let generator beat the discriminator
# having too many layers for the discriminator may not help.
# flatten layer
flat = tf.reshape(relu3, (-1, 4*4*512))
# logits and outputs
# the output node size should be 1
# because discriminator tries to predict whether the image is real or fake (1 or 0)
logits = tf.layers.dense(flat, 1, kernel_initializer=kinit)
outputs = tf.sigmoid(logits)
return outputs, logits
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)
Generator¶
Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.
def generator(z, out_channel_dim, is_train=True):
"""
Create the generator network
:param z: Input z
:param out_channel_dim: The number of channels in the output image
:param is_train: Boolean if generator is being used for training
:return: The tensor output of the generator
"""
# TODO: Implement Function
kinit = tf.random_normal_initializer(stddev=0.02)
# variables should be reused when not training
with tf.variable_scope("generator", reuse=not is_train):
# fully connected layer as the first layer
# to get 7x7 shape via 'VALID' padding
# this layer starts with 2x2 shape rather than 4x4
# I also applied large number of filters for generator
# in order for generator to beat discriminator
x1 = tf.layers.dense(z, 2*2*1024, kernel_initializer=kinit)
x1 = tf.reshape(x1, (-1, 2, 2, 1024))
bn1 = tf.layers.batch_normalization(x1, training=is_train)
# ReLU for generator layers
# as suggested in the DCGAN paper
relu1 = tf.maximum(0.0, bn1)
# now the shape is 2x2x1024
# layer2
# when padding's 'VALID',
# next_shape = ceil((shape - kernel_size + 1) / strides)
# to get 7x7 from 2x2,
# 2 = ceil((7 - 5 + 1) / 2)
# 2 = ceil((8 - 5 + 1) / 2)
# tensorflow chooses the odd value (7 among 7 and 8) when
# applying conv2d_tranpose with 'VALID' padding
x2 = tf.layers.conv2d_transpose(relu1, 512, 5, strides=2,
padding='VALID', kernel_initializer=kinit)
bn2 = tf.layers.batch_normalization(x2, training=is_train)
relu2 = tf.maximum(0.0, bn2)
# now the shape is 7x7x512
# layer3
x3 = tf.layers.conv2d_transpose(relu2, 256, 5, strides=2,
padding='SAME', kernel_initializer=kinit)
bn3 = tf.layers.batch_normalization(x3, training=is_train)
relu3 = tf.maximum(0.0, bn3)
# now the shape is 14x14x256
# output layer
# should not apply batch normalization here
logits = tf.layers.conv2d_transpose(relu3, out_channel_dim, 5, strides=2,
padding='SAME', kernel_initializer=kinit)
# now the shape is 28x28xout_channel_dim
# tanh activation function will make the logits range from -1 to 1
outputs = tf.tanh(logits)
return outputs
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)
Loss¶
Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:
discriminator(images, reuse=False)generator(z, out_channel_dim, is_train=True)
def model_loss(input_real, input_z, out_channel_dim):
"""
Get the loss for the discriminator and generator
:param input_real: Images from the real dataset
:param input_z: Z input
:param out_channel_dim: The number of channels in the output image
:return: A tuple of (discriminator loss, generator loss)
"""
# TODO: Implement Function
# generator model
g_model = generator(input_z, out_channel_dim)
# discriminator with real input
d_model_real, d_logits_real = discriminator(input_real)
# discriminator with fake input
# real and fake models should have the same weights, so set reuse True
d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)
# real loss
# real data's labels should all be 1
# this is not suggested in the DCGAN paper, but
# I'll try to apply label smoothing as shown in the DLND lectures
# with 0.1 label smoothing (i.e multiplying by (1 - 0.1) which is 0.9)
d_loss_real = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real,
labels=tf.ones_like(d_model_real) * 0.9))
# and 0 for fake labels
d_loss_fake = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake,
labels=tf.zeros_like(d_model_fake)))
# sum of real and fake losses should be the discriminator loss
d_loss = d_loss_real + d_loss_fake
# now the generator loss
# this should be the opposite to d_loss_fake
# because generator wants to fool the discriminator
# also applied label smoothing here as well
g_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake,
labels=tf.ones_like(d_model_fake) * 0.9))
return d_loss, g_loss
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)
Optimization¶
Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).
def model_opt(d_loss, g_loss, learning_rate, beta1):
"""
Get optimization operations
:param d_loss: Discriminator loss Tensor
:param g_loss: Generator loss Tensor
:param learning_rate: Learning Rate Placeholder
:param beta1: The exponential decay rate for the 1st moment in the optimizer
:return: A tuple of (discriminator training operation, generator training operation)
"""
# TODO: Implement Function
# get variables
t_vars = tf.trainable_variables()
d_vars = [var for var in t_vars if 'discriminator' in var.name]
g_vars = [var for var in t_vars if 'generator' in var.name]
# to update population statistics for batch normalization
# wrap the train operations with tf.control_dependencies
with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)
return d_train_opt, g_train_opt
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np
def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
"""
Show example output for the generator
:param sess: TensorFlow session
:param n_images: Number of Images to display
:param input_z: Input Z Tensor
:param out_channel_dim: The number of channels in the output image
:param image_mode: The mode to use for images ("RGB" or "L")
"""
cmap = None if image_mode == 'RGB' else 'gray'
z_dim = input_z.get_shape().as_list()[-1]
example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])
samples = sess.run(
generator(input_z, out_channel_dim, False),
feed_dict={input_z: example_z})
images_grid = helper.images_square_grid(samples, image_mode)
pyplot.imshow(images_grid, cmap=cmap)
pyplot.show()
Train¶
Implement train to build and train the GANs. Use the following functions you implemented:
model_inputs(image_width, image_height, image_channels, z_dim)model_loss(input_real, input_z, out_channel_dim)model_opt(d_loss, g_loss, learning_rate, beta1)
Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.
def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
"""
Train the GAN
:param epoch_count: Number of epochs
:param batch_size: Batch Size
:param z_dim: Z dimension
:param learning_rate: Learning Rate
:param beta1: The exponential decay rate for the 1st moment in the optimizer
:param get_batches: Function to get batches
:param data_shape: Shape of the data
:param data_image_mode: The image mode to use for images ("RGB" or "L")
"""
# TODO: Build Model
# lists to store losses while training
d_losses = []
g_losses = []
# get the parameters for the placeholder function
# width and height
image_width = data_shape[1]
image_height = data_shape[2]
# color channels
if data_image_mode == "RGB":
image_channels = 3
elif data_image_mode == "L":
image_channels = 1
else:
return "data_image_mode argument should be either 'RGB' or 'L'"
# build placeholders
input_real, input_z, lr = model_inputs(image_width,
image_height,
image_channels,
z_dim)
# build losses
d_loss, g_loss = model_loss(input_real, input_z, image_channels)
# build optimizers
d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)
# initialize steps (show every 100 steps)
steps = 0
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for epoch_i in range(epoch_count):
for batch_images in get_batches(batch_size):
# TODO: Train Model
steps += 1
# batch image's values range from -0.5 to 0.5
# so we multiply them by 2
# to fit the tanh output which ranges from -1 to 1
batch_images = batch_images * 2.0
# random noise for generator
# the range should be the same as shown in
# show_generator_output() function
batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))
# run optimizers
_ = sess.run(d_opt, feed_dict={input_real:batch_images,
input_z:batch_z,
lr:learning_rate})
_ = sess.run(g_opt, feed_dict={input_real:batch_images,
input_z:batch_z,
lr:learning_rate})
# show generator output and loss every 100 steps
if steps % 100 == 0:
show_generator_output(sess, 25, input_z, image_channels, data_image_mode)
train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
train_loss_g = g_loss.eval({input_z: batch_z})
print("Discriminator Loss: {:.4f}...".format(train_loss_d),
"Generator Loss: {:.4f}".format(train_loss_g))
d_losses.append(train_loss_d)
g_losses.append(train_loss_g)
# plot learning curves
pyplot.plot(d_losses, label='d_loss')
pyplot.plot(g_losses, label='g_loss')
pyplot.legend()
pyplot.show()
MNIST¶
Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.
# these values are suggested in the DCGAN paper
# batch size of 128, learning rate of 0.0002
# and beta1 of 0.5
# After many tries,
# I noticed that training GANs is the matter of balancing
# the two networks
# I ended up reducing the batch size from 128 to 32
# and incresing the learning rate from 0.0002 to 0.002
batch_size = 32
z_dim = 100
learning_rate = 0.0003
beta1 = 0.5
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2
mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
mnist_dataset.shape, mnist_dataset.image_mode)
CelebA¶
Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.
# for face generation,
# I just tried the same parameters
batch_size = 32
z_dim = 100
learning_rate = 0.0003
beta1 = 0.5
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1
celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
celeba_dataset.shape, celeba_dataset.image_mode)
Submitting This Project¶
When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.